US 7214327 B2
A method and apparatus for dry etching pure Cu and Cu-containing layers (220, 310) for manufacturing integrated circuits. The invention uses a directional beam of O-atoms with high kinetic energy (340) to oxidize the Cu and Cu-containing layers, and etching reagents (370) that react with the oxidized Cu (360) to form volatile Cu-containing etch products (390). The invention allows for low-temperature, anisotropic etching of pure Cu and Cu-containing layers in accordance with a patterned hard mask or photoresist (230, 330).
1. A method for anisotropically etching a Cu-containing layer comprising the steps of:
providing a Cu-containing layer overlying a substrate;
introducing a directional beam of neutral oxygen atoms having a kinetic energy between 10 eV and 1 eV;
oxidizing the Cu-containing layer by exposure to the beam of oxygen atoms;
introducing a reagent capable of forming volatile etch products when reacted with the oxidized Cu-containing layer;
removing the etch products from the Cu-containing layer.
2. The method according to
3. The method according to
4. The method according to
5. The method according to
6. The method according to
7. The method according to
8. The method according to
This application claims priority to U.S. Provisional Application Ser. No. 60/392,045, filed Jun. 28, 2002; and is related to co-pending Provisional Application Ser. No. 60/379,754 filed on May 14, 2002, the entire contents of which are incorporated herein by reference.
The present invention relates to the field of semiconductor processing and more particularly to a method and apparatus for anisotropic etching and patterning of pure Cu and Cu-containing layers used in integrated circuits.
Copper (Cu) is emerging as the metal of choice in a wide variety of semiconductor applications. Lower electrical resistivity, coupled with improved electromigration performance and increased stress migration resistance are important material properties that favor the use of Cu over aluminum (Al) in interconnect lines and contacts. The lower electrical resistance is critical since it allows signals to move faster by reducing the RC time delay. The superior resistance to electromigration, a common reliability problem in Al lines, means that Cu can handle higher power densities. An equally important benefit of using Cu over Al is that the manufacturing cost for a Cu metallization scheme can be lower due to new processing methods that reduce the number of manufacturing steps and alleviate the need for some of the most difficult steps.
The capability to process substrates anisotropically permits the production of integrated circuit features at precisely defined locations with sidewalls that are essentially perpendicular to the surface of a masked overlayer. The introduction of Cu into multilevel metallization architecture requires new processing methods for Cu patterning. Because Cu is difficult to dry etch, new process schemes have been developed for Cu patterning.
The damascene approach is based on etching features in the dielectric material, filling them with Cu metal, and planarizing the top surface by chemical mechanical polishing (CMP). Dual damascene schemes integrate both the contacts and the interconnect lines into a single processing scheme. However, Cu CMP technology is challenging and it has difficulty defining extremely fine features. In addition, CMP suffers from yield-detracting problems of scratching, peeling, dishing and erosion.
An alternative to the damascene approach is patterned etching of a Cu layer. The patterned etch process involves deposition of a Cu layer on a substrate, the use of a patterned hard mask or photoresist over the Cu-containing layer, patterned etching of the Cu layer using a reactive ion etching (RIE) process; and deposition of dielectric material over the patterned Cu-containing layer. Patterned etching of Cu can have advantages over damascene processes since it is easier to etch fine Cu patterns and then deposit a dielectric layer onto the Cu pattern, than it is to get barrier layer materials and Cu metal to adequately fill small feature openings in a dielectric film.
Patterned etching using RIE processes is characterized by relatively low pressures and high ion bombardment energies. While ion bombardment is needed to achieve the desired degree of etch anisotropy, it is also responsible for secondary damage to the underlying microstructure. As semiconductor devices have become increasingly more integrated and new advanced materials, such as copper and low-k dielectric materials, have been introduced to improve the circuit properties, the damage caused by fabrication processes presents an increasingly serious problem. A leading reason for damage formation is essentially the incidence of energetic particles, such as ions and UV photons, from the plasma environment to the substrate surface.
The primary etch reagent for removing Cu layers is traditionally a chorine-containing gas in gas mixture that includes argon (Ar). Removal of pure Cu layers and Cu-containing layers with high Cu-content using chlorine plasma essentially involves physical sputtering of the low-volatility CuClx surface layer by energetic ions in the plasma. The Cu removal rates are very low when using this method and another drawback is that the sputtered CuClx coats the chamber walls and this requires periodic cleaning of the etching chamber. An equally serious problem is encountered when high-aspect-ratio features are etched in chlorine plasma and the sputtered CuClx products redeposit on the feature sidewalls where the effects of physical sputtering are reduced.
When the abovementioned chlorine-based etching process is carried out at elevated temperatures (>200° C.) to increase the volatility of the reacted Cu-containing layer, corrosion can occur due to the accumulated CuClx etch residues on the surface. If these residues are not removed by a post-etch cleaning step, they can cause continuing corrosion of the Cu even after the application of a protective layer over the etched features.
Other energy sources have been suggested to increase the etching rate. These approaches include exposing the etching surface to UV or IR light sources to accelerate desorption of CuClx from the etching surface. However, these approaches are not practical for semiconductor batch processing of large substrates due to poor etch uniformity, high cost and added equipment complexity, and reliability problems.
Accordingly, it is desirable to develop a method for anisotropic etching in semiconductor manufacturing using plasmas that are essentially free of energetic ions and photon particles. Such particles are ordinarily required to achieve an isotropic etch with atomic oxygen plasmas in the prior art, but cause secondary damage to the underlying substrate. Furthermore, the reaction products should be highly volatile and easily removed from the etched substrate. In addition, the aforementioned limitations that are encountered when etching Cu-containing layers using conventional chlorine chemistry, show that there is a need for new low temperature dry etching methods in semiconductor manufacturing using chemical approaches that do not involve chlorine-based reactants.
It is a primary object of the present invention to provide a method and an apparatus for anisotropic dry etching of pure Cu layers and Cu-containing layers used in integrated circuits.
The above and other objects are achieved, according to the present invention, by providing a method and an apparatus that uses a directional beam of high kinetic energy (hyperthermal) neutral O-atoms to anisotropically oxidize Cu layers, and an etching gas comprising an etch reagent that readily reacts with the oxidized Cu layers to form volatile Cu-containing etch products. The high directionality of the O-atom beam allows for anisotropic etching of the Cu layers in accordance with a patterned hard mask or photoresist.
A more complete appreciation of the invention and many of the attendant advantages thereof will become readily apparent with reference to the following detailed description, particularly when considered in conjunction with the accompanying drawings, in which:
In general, the present invention pertains to a method and an apparatus for etching pure Cu and Cu-containing layers in manufacturing integrated circuits. The method uses a directional beam of high kinetic energy O-atoms to oxidize the Cu layers, and an etch reagent that forms volatile Cu-containing etch products when reacted with the oxidized Cu layers.
Importantly, the etch reagent reacts very slowly with unoxidized Cu on feature sidewalls, but reacts readily with oxidized Cu surfaces at low temperatures. Therefore, in order to achieve anisotropic etching and avoid undercutting of etch features, the oxidation rate of Cu on the vertical sidewalls of the etch features must be substantially slower than the oxidation of Cu on the horizontal surfaces. This is achieved using a directional O-atom beam.
The directional beam is preferably a hyperthermal beam of neutral O-atoms that has high kinetic energy and can be produced with low divergence from the substrate normal, thereby allowing production of masked structures with reduced ion damage and charge build-up problems often encountered in plasma processes.
In general, hyperthermal atomic beams contain neutral atoms that have high kinetic energy (Ek≧1 eV) compared to thermalized atoms (Ek˜0.05 eV) produced by conventional glow discharge devices, such as plasma ashers and sputter etching apparatus. The kinetic energy of hyperthermal atomic beams can be has high as several hundred eV.
The use of hyperthermal O-atom beams to oxidize and etch films has been shown to result in reduced ion bombardment damage and charge-induced damage that is commonly observed during plasma etching processes, such as reactive ion etching (RIE). For example, hyperthermal O-atoms have been used for anisotropic etching in semiconductor manufacturing, particularly in the areas of photoresist stripping and multilayer lithography applications, with little or no secondary damage to the underlying substrate.
Koontz and Cross in U.S. Pat. No. 5,271,800 entitled “Method for anisotropic etching in the manufacture of semiconductor devices,” disclose a method for anisotropic etching of a hydrocarbon polymer coating using a beam of hyperthermal atomic oxygen, to generate a desired pattern on a semiconductor wafer. The hyperthermal beam has kinetic energy in the range between about 0.2 eV to 20 eV and is produced using a continous optical discharge laser.
In the current invention, the use of a hyperthermal beam of neutral O-atoms to oxidize Cu layers and Cu-containing layers and subsequent reaction of the oxidized layers with etch reagents to form Cu-containing etch products, allows for anisotropic etching at low substrate temperatures. In one embodiment, the hyperthermal beam of O-atoms is generated in a separate chamber from the processing (etching) chamber and the beam contains only neutral O-atoms when it enters the processing chamber. Alternatively, the O-atom beam is generated inside the processing chamber. The etch reagent is introduced into the processing chamber using effusive sources and is therefore not exposed to a plasma source that could result in deleterious decomposition of the etch reagent prior to reaction with the oxidized Cu layers.
The neutral O-atoms in a hyperthermal beam are schematically represented by arrows 340 in
In contrast, when the etch reagent 370 adsorbs onto the oxidized Cu layer 360 forming the adsorbed etch reagent 380, it reacts spontaneously, even at low temperature, with the oxidized Cu layer. In addition to exposure to the etch reagent 370, the horizontal surfaces 360 are under constant energetic O-atom bombardment 340, which, in addition to oxidizing the surface, further aids desorption of etch products 390 from the surface and allows etching of Cu-containing layers at low temperatures.
The kinetic energy of the O-atoms affects the oxidation of the horizontal surfaces. When O-atoms with high kinetic energy interact with the horizontal surfaces, the O-atoms penetrate the surface layer, loose their kinetic energy (become thermalized) as they become embedded in the surface layer, and thereby oxidize the surface layer. If the kinetic energy is below a certain threshold energy, the O-atoms are more likely to “bounce back” without becoming thermalized, and this can result in undesired sidewall oxidation and sidewall etching.
Process conditions that enable the desired etching of the Cu-containing layer in the current invention may be determined by direct experimentation and/or design of experiments (DOE). For example, adjustable process parameters can comprise the kinetic energy of the O-atoms in the hyperthermal beam, substrate temperature, process pressure, choice of process gases and relative gas flows of the process gases.
The etch reagent 370 can comprise chemical compounds such as β-diketones. This family of compounds is characterized by a molecular structure involving two carbonyl functional groups that are separated by a single carbon atom to which is attached at least a single hydrogen atom. Therefore, β-diketones are also characterized by keto-enol tautomerism in which the enol form is stabilized by conjugated double bonds. Tautomerism refers to an equilibrium between two different structures of the same compound.
In addition, the conjugated nature of the double bonding implies significant delocalization of π-electrons which affects the ability to bond to metal atoms. These delocalized electrons readily interact with the d-electrons of metal atoms to form coordination compounds. Due to the presence of the two carbonyl groups, each β-diketonate ligand coordinates twice with the metal atom to form a closed ring or chelated structure.
Common β-diketone compounds include acetylacetone (acacH, where R1═R2═CH3); 1,1,1-trifluoro-acetylacetone (tfacH, where R1═CF3 and R2═CH3); and 1,1,1,5,5,5-hexafluoro-acetylacetone (hfacH, where R1═R2═CF3). These compounds are commercially available and are widely used in chemical synthesis. The well known compound hfacH is used here as an example to illustrate chemical properties of β-diketones and the overall surface etching reactions of a β-diketone with an oxidized Cu (CuO and/or Cu2O) layer.
HfacH has the highest vapor pressure (85 Torr at 20° C.) of the above β-diketones, and this allows for easy transport of the gaseous hfacH etching gas to the processing chamber using a delivery system that can comprise a bubbler system and a mass flow controller (MFC). The bubbler system can be used with or without a carrier gas such as argon (Ar). When a carrier gas is used, it is bubbled through the hfacH liquid and becomes saturated with the hfacH vapor. The partial pressure of the hfacH vapor in the process chamber is controlled by the temperature of the hfacH liquid in the bubbler. Exemplary gas flow rates of hfacH and a carrier gas are less than 1000 sccm, preferably being less than 500 sccm. Alternatively, a liquid injection system can be used to deliver the hfacH to the processing chamber. The handling and use of etch reagents such as β-diketones reagents in this invention is well known in the art.
An important property of β-diketones is that they readily form stable anions (e.g., hfac) that can react with metal cations (e.g., Cu, Pd, and Pt ions) to form stable complexes that have relatively high volatility. For example, Cu(hfac)2 and Cu(hfac)VTMS precursors comounds are used for chemical vapor deposition (CVD) of Cu layers.
The net surface reactions between hfacH and oxidized Cu can be written as:
An inert gas can be added to any one of the aforementioned process gas chemistries. The inert gas may include at least one of argon, helium, krypton, xenon, and nitrogen. For example, the addition of inert gas to the process chemistry is used to dilute the process gas or adjust the process gas partial pressure(s).
Methods of reacting β-diketones with oxidized copper layers for in-situ cleaning of processing chambers have been disclosed. These methods involve oxidation of deposited residues and subsequent removal of the oxide residues from the chamber, thereby eliminating the need to disassemble the chamber for cleaning.
Nguyen and Chameski in U.S. Pat. No. 6,284,052 entitled “In-situ method of cleaning a metal-organic chemical vapor deposition chamber,” describe a method to clean the interior surfaces, and especially the chuck, of a metal deposition chamber. The method first oxidizes the surface to be cleaned with an oxygen plasma, and then removes the oxide products as a vapor with the use of hfacH.
Koide et al. in U.S. Pat. No. 5,993,679052 entitled “Method of cleaning metallic films built up within thin film deposition chamber,” describe a method that includes an oxidation step to oxidize a metallic film, a complexing step to complex the oxide film, and a sublimations step to sublimate the complex. The conditions of these cleaning steps are set so that the oxidations step is the rate determining step.
The focus of the current invention is to enable isotropic etching of Cu layers and Cu-containing layers for patterning multilayer structures. The main purpose of the above-mentioned methods is to clean and remove metallic films from processing chambers through the formation of volatile etch products from the reaction of oxidized metal films with β-diketones.
Substrate 620 is transferred into and out of process chamber 605 through a slot valve (not shown) and chamber feed-through (not shown) via robotic substrate transfer system where it is received by substrate lift pins (not shown) housed within substrate holder 615 and mechanically translated by devices housed therein. Once the substrate 620 is received from the substrate transfer system, it is lowered to an upper surface of the substrate holder 615.
In an alternate embodiment, the substrate 620 is affixed to the substrate holder 615 via an electrostatic clamp (not shown). Furthermore, the substrate holder 615 further includes a cooling system including a re-circulating coolant flow that receives heat from the substrate holder 615 and transfers heat to a heat exchanger system (not shown), or when heating, transfers heat from the heat exchanger system. Moreover, gas may be delivered to the backside of the substrate to improve the gas-gap thermal conductance between the substrate 620 and the substrate holder 615. Such a system is utilized when temperature control of the substrate is required at elevated or reduced temperatures. Vacuum pump system 635 preferably includes a turbo-molecular vacuum pump (TMP) capable of a pumping speed up to 5000 liters per second (and greater) and a gate valve for throttling the chamber pressure.
A controller 645 includes a microprocessor, a memory, and a digital I/O port capable of generating control voltages sufficient to communicate and activate inputs to the processing system 600 as well as monitor outputs from the processing system 600. Moreover, the controller 645 is coupled to and exchanges information with the process chamber 605, the O-atom beam apparatus 610, the process monitor system 650, the gas injection system 625 and the vacuum pump system 635. A program stored in the memory is utilized to control the aforementioned components of a processing system 600 according to a stored process recipe. One example of controller 645 is a digital signal processor (DSP), model number TMS320, available from Texas Instruments, Dallas, Tex.
The process monitor system 650 can comprise, for example, a mass spectrometer system to measure gaseous species, such as etch reagents and etch by-products in the processing environment. In general, the process monitor system 550 is a versatile diagnostic tool capable of performing multiple tasks such as process analysis and endpoint detection. The process monitor system 650, shown in
Examples of hyperthermal O-atom sources will now be described. Various methods and apparatus have been developed for generating a beam of neutral hyperthermal O-atoms. The dissociative ionization of molecular oxygen (O2) supply gas to produce atomic oxygen can be carried out using optical laser-induced discharges, dc- or ac arcs, inductive coupling, microwave, or electron bombardment discharges, that are usually in a separate chamber connected to the wafer processing chamber. O-ions that are formed from the dissociation of O2 in the plasma are accelerated to the selected high kinetic energy and then charge neutralized, forming a beam of neutral hyperthermal O-atoms. The charge neutralization of the O-ions can be carried out using a variety of different method. An example is electron transfer from interaction of O-ions with metal surfaces at grazing incidence.
Chen and Yvonne in U.S. Pat. No. 6,331,701 entitled “RF-grounded sub-Debye neutralizer grid,” describe apparatus for generating a hyperthermal beam of neutral O-atoms with kinetic energy ranging from 20 eV to 400 eV and a beam diameter of 1 inch to greater than 10 inches for processing large substrates. The hyperthermal beam is generated by passing accelerated oxygen ions from a RF-plasma through a neutralizer grid where the oxygen ions are charge neutralized through forward surface scattering. The hyperthermal beam has high flux of neutral O-atoms and a small divergent angle (˜6–3°).
Optical laser-induced discharges can employ pulsed laser radiation as an energy source for dissociation the supply gas (e.g., O2) in an expansion nozzle. The expansion creates a hyperthermal beam of neutral reactive species (e.g., O-atoms) that can be collimated using an orifice between the expansion chamber and the processing chamber.
It should be understood that various modifications and variations of the present invention may be employed in practicing the invention. It is therefore to be understood that, within the scope of the appended claims, the invention may be practiced otherwise than as specifically described herein.